Dongbu Corporation has begun deploying a smart safety management system that pairs unmanned drones with AI video analysis across three major construction sites. The move targets faster hazard detection and a reduction in reliance on manual patrols, directly addressing the high injury rates that continue to plague large-scale projects.
The system, developed with spatial intelligence firm Meissa, is now active at the OTOKI Logistics Center in Hwaseong, the Jinhae New Port dredged soil disposal site, and a private-public housing project in Namyangju. Dongbu tailored the drone station operations and AI models to the specific work environment and risk profile of each location.
How the unmanned stations work
Each site uses a drone station where aircraft take off, fly pre-set routes, and land without an on-site pilot. The drones inspect expansive areas and hard-to-access zones on a fixed schedule. Dongbu expects the automated patrols to sharply cut the time safety managers spend walking perimeters, freeing them for direct oversight of high-risk tasks.
Footage streams into a dedicated AI platform. The software automatically flags workers entering heavy equipment zones and tracks the real-time location of personnel tagged as high-risk, giving supervisors a live map for emergency response.
AI catches what humans miss
The video analysis also scans for common safety violations that lead to fines and injuries. The model recognizes missing hard hats, absent safety vests, and unfastened harnesses. Alerts go straight to on-site managers, creating a record that links each violation to a specific time and location.
Beyond safety, the daily drone imagery is stitched into 3D site models to visualize construction progress. Project teams use this data in coordination meetings to spot clashes between subcontractors and to identify hazards embedded in the next phase of work. The system also pulls major risk factors from upcoming tasks and pushes customized daily safety briefings to workers via large screens and mobile devices.
Rollout plans
A Dongbu Corporation official said the system "is receiving great responses as a practical safety management tool that goes beyond simple monitoring to prevent safety accidents and improve construction quality." The company will mandate the technology on projects with construction costs of KRW 200 billion or more starting in 2027. The target is full deployment across all sites by 2029.
Why this matters for real estate and construction professionals
Dongbu's internal mandate signals that drone-and-AI safety systems are moving from pilot experiments to contractual requirements on large projects. For developers and general contractors, the 2027 threshold creates a clear timeline for budgeting this technology into project overhead. For safety engineers, the shift demands new skills in configuring automated inspection routes and interpreting AI-generated violation logs-a competency gap that training paths like the AI Learning Path for Safety Engineers are designed to close. As owners and insurers increasingly reward automated oversight with lower premiums, firms that delay adoption risk losing bids on major private and public works.
Your membership also unlocks:
